How to debug email deliverability

[00:00:00] Mansi: Hey, everyone. Welcome to the growth chat by Mailmodo. This is Mansi, your host for the evening and a marketer at Mailmodo. Today, we are going to talk about email deliverability and the various factors that affect it and how we can fix it. Who better than two deliverability gigs who spend their day making sure that emails get delivered and building strategies around it. Sadikshya Pant from Mo Engage and Surya from Mailmodo are here with us today to share their frameworks and strategies to help you get your emails. Let's hear from you guys about yourself. Sure.

[00:00:36] Sadikshya: Hi everyone. I'm Sadikshya. I am currently working with MoEngage. I, I actually entered the deliverability field straight off during my Vick, during my internship, I got into the email deliverability team and then I continued. So it's been seven years of learning and growing and helping customers solve it, and deliver issues. [00:00:55] Surya: Thank you. First of all, for having me on board.

Hi, I'm Surya. I've been in the [00:01:00] industry for the last six plus years, helping clients solve their deliverability problems and also fixing their email strategies. Also helping them with better design and copies besides being an email geek. I'm also a cinephile and, I like cricket.

[00:01:15] Mansi: Awesome. . Great introduction guys. Thank you. So we'll just move on. And so I, I. So, you know, we have seen like a lot of people confuse delivery with deliverability and there's so can you both share your definitions of email deliverability and how it impacts revenue?

[00:01:34] Sadikshya: Sure. So when we say delivery, what does it mean that you send an email and the recipient server accepted your email. So it's basically when you send out to hundred.com customers, it just, you know, if 99 of those emails were accepted by the Gmail server, you have 99% delivery rates, but deliverability means how many of those got into your inbox. So when you create an email campaign, right, it has an [00:02:00] object. If you have a welcome campaign, you wanna onboard your customer to the brand. You wanna let them know what they have signed up into. If you send an abandon card campaign, you wanna nudge the people who have added something to the card, too.

But if your emails go to your spam folder, if they don't, if the customer isn't able to see that, so that doesn't fulfil the purpose of the campaign, per se, that's just, I would say deliverability is very, very important because it's literally, it's a major of how many of the people got their email, where they should get it in the first place. Okay. So, yeah.

[00:02:39] Surya: Right, right. Agree with you Sadiskhya. And I always like to draw a real-life analogy and as Mansi just mentioned how deliverability depends on the ROI. It, so I always compare this to the case of a travelling salesman. So for a travelling salesman, the first challenge is to gain entry into the door and [00:03:00] then he's able to, he, or she's able to sell a product, right. Similarly, with emails, it's all about first getting the first step. Is getting the email in the inbox, right. And then comes the product then comes, what kind of copy you have written then comes what kind of design you have created with the best design, with the best copy. If you're not inboxing, your ROI will simply go for a do.

So you must be following all the best practices of deliverability, which we are going to discuss in the course of this session. So, yeah, ROI and deliverability are very closely interlinked, which a lot of marketers tend to neglect at times, but it is, is a very important aspect of getting the optimum conversion rate from the emails that you're sending an email being one of the highest ROI generating channels.

You must be, it's literally like I, again, draw an analogy. It's similar to [00:04:00] a cricket captain. Using his best polar most economically, right. Using his best polar at the right time so that he's able to save energy. He's able to take S so, yeah, deliverability and otherwise are very closely interlinked.

[00:04:19] Sadikshya: And adding to that, I think I wanna bust myth saying, you know when we talk to marketers, they it's not their field. So deliverability is a so subject for them deliver to people can be seen sometimes that someone who tells them, no, you need three weeks of ramp up. No, you can't spend this much. You know, know you can't blast your emails, but at the end of the day for every deliverability person, our end goal is to find that sweet spot where, where you can reach out to the maximum number of customers optimally.

Right? Sure. We can blast two emails to everybody, but if you know, more than half of them are getting it into spam. Then your ma then your campaign isn't working. I mean, it's not working [00:05:00] as good as it should be working. Right. So that's why deliverability plays a very important role in terms of what you get out of your email program as a whole. So as a delivery person, I would say what I'm trying to achieve the end of the day is that you can reach out to post numbers for your customers if possible, all the people in your database, as many times as possible without impacting your reputation, per se say. So finding that optimal spot where you are neither over mailing the campaigns, as well as needed under mailing is what essentially thet is all about.

Right? Right. And while monitoring deliverability, you have to pretty much be a span. Filterer you have to think like a span. You need to understand how span filter work. Right. You have to understand deliverability is never a problem. With the ESP you are sending emails with, it's always a problem with the strategy you have incorporated. It's a problem with the [00:06:00] ESP. Obviously, if you know, the infrastructure is not being maintained, rightly, which is a very rare case. And later on, we are going to discuss the factors which deliver duty depend on. But I would say you know, you, you have to, as a deliverability person, you have to always keep your eyes open for any red flag in email strategy or content.

[00:06:20] Mansi: Awesome. So alright, so we all right. So we have very, very defined deliverability and obviously it's impact on the revenue. So let's look into the factors, as you mentioned, Surya let's, let's look into factors that affect deliverability.

[00:06:31] Sadikshya: If I have to say what effects the, I really wanna say everything does okay. But if I have to point out two, three major factors, then first I would say it would be the setup, right setup, and then how you collect your audience and how you are marrying your content with your cohorts. And those are the basic pillars, which upon which your email program is built upon. When we say authentication, it means Mansi, can you share this please? When we say authentication, it means the, [00:07:00] like how well your domain and IPS are authenticated, right? Like whether your IP and domain are eligible to send emails on your behalf. I mean, if you look at this example can go to the first slide. So you see email from one is.io. and if you look at the mail by and signed by here, it's alert, moge.io and mortgage.io, and DKI is signed by moge.io. So it's basically your SPF F DKI IM your from domain. All of them are the same domain. It looks pretty branded. And then there's your if your lead brand also said something like the, you know, if your links start with something like link dot, more engaged.io and your, your IP has IDs going, something like, you know, PTR dot, more engaged.io, it looks like everything in your, everything in your setup is branded.

Now can you go through the second slide? Yeah. Now, if you look at this example, you see an email from [00:08:00] brand.com and it did come to the inbox, but it came via Moin IO. So it's basically you leveraging the reputation, leveraging the authentications of mo.io to go into this, to basical. Go out right. And click on the next slide. Similar example, it's the same brand.com, but it's going by a send.net here again. It's utilizing the authentications of your send.net. So while both of them are utilizing different domains, it, the reputation, your reputation of your domain is not upon brand.com. It's basically leveraging the reputation of a third party, right?

Who engage at IO, send.net. So depending upon how many other brands are using the send grid.net or more engaged IO and how it's performing your email campaign will will, will, will get affected. It might either be good if you have all the good senders in there, or it might be bad if somebody else is cruiser.

Right. Can you go to the next slide now, when we talk about D [00:09:00] a or D a is like a very simple exam solution to prevent your email address from getting. Like if here you can see their brand.com doesn't have any demark in place. Can I go to next slide here while Mo engaged IO has a DMAR, which has policy, which says equals to, right? Which means if your DMAR doesn't pass, if your emails fail basic, basic or authentications, then your emails will get rejected. Can anyone mix like this? Yeah. So instead of emails going, you know, via somebody, something else, like, you know, some, something else to mean.com, like we saw brand.com via Mo engaged do IU or somebody.net.

We'd get an error message. So if a Stamer is trying to use your domain to send spa emails, they wouldn't be able to send it. That's why authentication is very, very important. It's very imperative that you [00:10:00] have proper authentications that are for all your domains and sub domains and you, and you. Strict teamwork in place.

So that spams can't spoof your domain and can't imperson. And apart from authentications, how you collect your we can stop sharing. You it's very important how you collect your audiences, right? If your audience collection mechanism is flawed from the very beginning, for example, if you don't, if you are getting audiences from a third party or by an affiliate brand, or if they basically, if your users haven't specifically meant uh, signed up for your brand, then it means they're not interested on the brand or they, they may not even have heard of the brand says, no matter how, how good or content you send to them, they're not going to engage positively with you.

So if your basic basic authentication is flaw, if your basic. Authentication is flawed. Then others can impersonate you and your how [00:11:00] your reputation might not even be built upon what you are doing. It might be affected by what somebody else is doing at the same time. If your audience collection methods are insecure, if they're in good enough, then essentially you are not sending you emails to the right person who might be interested in what you have to say at such. It's not going to work very well, no matter how good your content is. Similarly it go. When I say marrying the content with the cohorts, it means as a brand, you mostly as a brand, you might not have just one product that you wanna sell. Right? You may have different options, different features or different products.

So at, since you have to figure out who, which cohort of people like which group of people wants exactly what right. That's and if you send them exactly what they're looking for, then. possibility of their engagement as well as their conversion is quite high. So yeah, if I have to point of just three, then I'd say [00:12:00] how you authenticate, like how you set up your infrastructure, how you collect your list as well as how will you marry your with the cohorts is very essential. So yeah, to you.

[00:12:12] Surya: Thank you. And I'm pretty much on the same page with what you just said. When I when I assess brands, I always follow a model, which I call the three R model the right database, the right infrastructure and the right content. Okay. So these, these three, as you mentioned, are very imperative to follow. So when it comes to right database, as you just mentioned, the database should be collected in a very a sophisticated way. You should not purchase a database. And many marketers out of the opinion that you can purchase database and just run the database through any cleansing tool, but that doesn't filter out span reps that doesn't filter out users who don't have the required affinity to your brand then comes the right infrastructure.

As you mentioned, [00:13:00] authentication is key and also maintaining your IPS, ensuring your IPS are not hitting out of spam reps, monitoring your IPS, ensuring they're not listed in prominent DSBS like spam house. And spam cup is very essential for ANP to follow, which we constantly do at main model. The third thing is the content reputation, content reputation is very important because all the spam filters are a IML based.

They learn and evolve. So if you have a phone number in the photo and you have sent. Sent lot of emails with that phone number in the photo to folks who are not interested and they have reported simply opened the emails and reported to spam. In that case, the phone number gets cashed. Similarly, there are you know, like best practices to follow in terms of content like not having Bitly links, which many marketers you know, are not aware of.

I have encountered with a lot of brands using bit links in the [00:14:00] content, which results in the emails, landing in spam for Gmail and many other ISPs, also maps map tags are a strict note for Gmails Pam filters. Sometimes all time title tags. Also. I recommend not to have that's mainly because, you know, marketers come from the lens of SEO. They think if they can boost their SEO rankings with all tags, they should do player similar game with emails, but it doesn't work like that because There are a lot of affiliate marketers in the industry. And sometimes it so happens that if you are assigning all equal to Facebook, to a Facebook item also an affiliate marketer is doing the same.

And as a result, what happens is all equal to Facebook, gets cashed, negatively affiliate marketing is sending emails to folks who have not subscribed to your database. They end up reporting to spam and your all tag get cashed. So it's best not to have all tax and title tax. And in the beginning, as I mentioned, right, we need to understand how spam filters work.

Spam filters [00:15:00] basically assign a certain score to your entire email. And if the score is beyond a certain threshold, the email ends up landing in inbox. If the score is below the threshold, the email lands in span. So you have to ensure that the overall score is more. So if, because of all tag, your score is coming down, you should get rid of all types because of score is coming down. You need to get rid of the bit links. Basically your content should not have anchored a lot of spa complaints in the past. If it has, and you are giving a fresh start to your email program, go for a, a different layout altogether, change the layout completely. For example, if you had a Z exact layout before go for an inverted pyramid layout, and that will solve your problem, but also make sure you are not committing the same mistake as before you are targeting the right user with the right content at the right time. And when it comes to engagement, I always say marry the platform activity with email activity. [00:16:00] Don't just target email active users, because maybe they're opening the emails today, but you never know when they will open the email and report to for that. You need to essentially track the brand affinity each and every user has.

So if you ask me for any brand, that would be four types of cohort. One cohort would, would be people who are active on email active on the platform. Second cohort would be people who are active on email, but are not active on the platform. Third cohort would be people who are not active on emails, but they're very active on the platform. And the fourth cohort is on the verge of churn users who are not active on email, active on platform. So you need to treat each and every cohort differently. You can treat them the same way. So for cohort one, which is email active platform active, I would recommend you send your from the D two C space, right?

From the D two C lens. You send your site wide offers to them. You can send your [00:17:00] regular newsletters to them. You can follow pretty much follow your campaign, calendar, email, active platform, and active. If you're sending some offers, encourage them to come to the platform by giving a slightly different offer. If you're throwing in a 20% discount book platform, active email, Throw in our 25% discount, maybe when it comes to email inactive platform, active, use your other channels to encourage them to open their emails. So I know of this brand which converted users to the first cohort, which is email active platform, active by sending push notifications and integrating them that look, we have sent you a voucher specialized for you to your inbox, open your email, to avail of the offer and the last cohort, which is email inactive platform, inactive, engage with them the least frequently.

And when you're sending them campaigns ensure that you're sending them BOGO offers. You have thrown a 20% discount to the first cohort, 25 to the second [00:18:00] maybe throw in a 60% of when you're sending, this is one of campaign once a month to these users to encourage them to, you know, come back to the platform. So this is a strategy I. And you know, like uh, and this is a strategy which if any of your follows will never go wrong, I assure you. [00:18:21] Sadikshya: Yeah. I agree with you. And strategy overall is fine. And then you have to test and you have to optimize and you have to see when you say email active and platform active, you have to add the, you know, X number of times in X number of days there.

So that is going be value for each user and yet, so that's what you have. Yeah. Recency frequency. You can have the recency frequency and see, you know, how many customers are your loyal customers. How many of them are about to sleep about to churn, about to, you know, about to ate and all that. And the other advantage off topic, sorry.[00:19:00]

oh yeah. I was going to say the other advantage of using the platform data. Is, you know, with the apple privacy really now the open metrics aren't as accurate as they used to be. It's not not saying open aren't reliable anymore. They're just not as reliable as they used to be. They're still open and clicks are still something you still need to monitor, but you also have to put in other metrics in there, like you have to look into other metrics to figure out which ones are active and which ones are inactive. So that's where your platform you know, the platform active and platform interaction really comes into the picture. And personally speaking, when apple private really came to the picture, when I figure out how we chase things, I realize for most of the brands I was already using the platform active you know, events and all.

And since we didn't have to make all those brands, didn't have to make all that many changes to not lose out on the active customers, because the open when coming through. Right. So when, once the changes come into picture, instead of [00:20:00] trying to find ways around it, we. You know, implement that as a best practice and follower such. One key thing about deliver is you need to be constantly evolving. You need to constantly find out things which would, you know which is going to be helpful in the future saying agree.

[00:20:17] Surya: Absolutely. So I was just going to go a little off topic and it's, it's unavoidable when to email each talk to go talk. So what I was asking you was what's your take on primary in boxing in India, at least, you know, like a lot of marketers think primary inboxing is the only way to get revenue. It's it's mainly because in India, people don't follow the concept of zero inbox. In the us people follow zero inbox wherein they check all their emails, delete unnecessary ones, and their mailbox is only filled with necessary emails.

So they end up checking promotions as well. The tendency to check promotions is much higher in the. As compared to India in [00:21:00] India, you can drive eyeballs only with primary and boxing. So what's your take on that? [00:21:06] Sadikshya: I, I would say for primary in boxing is great. Obviously, if you, if your promotions get into primary folder, it's obviously a boost, right?

But there's two ways to it. One is getting there because your users actually think that what you are sending, like the promotions that they're sending are worth being in the primary, right? That's what that happens. When you don't put an extra effort into forcing emails, into primary, you just follow your best practices. You send out the right audience, your frequencies tailored for your active, less active in active users, you know, for different cohorts and it's, and your content is what your users are users are looking for. So if all of, if you do everything right, then your emails might still go to PRI primary. I mean, I have lots of customers whose emails are going completed to primary in. When they're sending a hundred percent promotional content content. Right. But that didn't happen because [00:22:00] they, where you know, that that was their end, that wasn't their end goal. They were just following best practices. And slowly, slowly on their own Gmail started piecing those emails into primary because that's how the users reacted.

And when that happened, that didn't affect their reputation. Like they didn't get you know, they didn't get very inflated complaints. They did their unsubscribe. So still what they used to be. So they got there getting there by following best practices on a very organic way. I, that that's great. You know, that means you are doing something absolutely. Right. But following, you know, bricks and CLE X to push it there, that I, I would very strongly discourage discourage that. [00:22:44] Surya: Right. I always tell my marketers that grammar inboxing is a hack. It's not a fix. Yeah. It's a hack, right. It's a. And for all those people who want their emails to land in primary, one of the things which helps is having a lot of marketing automation, campaigns, [00:23:00] having a lot of transactional campaigns flowing from your domain because Gmail looks at the average open rate of your domain name.

And if you have very high generating open rate campaigns, then it while that's not the only way to inbox in primary, it's one of the ways to get there.

Yeah, yeah, yeah. Sorry for the diversion. Oh, that's fine. Say you can. Awesome. [00:23:23] Mansi: Thank, thank you guys. Cool. So, you know, with this vast experience that you guys have and like you've had helped so many companies fix their deliverability issues, can you share your frameworks and best practices for identifying and debugging problems?

[00:23:42] Sadikshya: Sure. When I Debu problems, the first thing I always look for is the authentications that's, you know, check if your all authentications are fine. And that means your SPFT IM D mark, your IP, your, you know, link branding, domain, your just basically all the authentications, as well as your [00:24:00] infrastructure. Like see if any of processes broken somewhere. Like sometimes your emails might go to spam because of process broke somewhere, not necessarily your fault, but again, know an issue. And at the same time sometimes suppose you're using TTP links, but your, the website implemented HSTS now you, your link should be TTPs S if you didn't do it, you know, at the same time, then your campaigns might start to get affected where not because now your links aren't working. So first things first, it would be due check the check, the authentications, you know, check if the setup is completely right. And after that, the other thing I would check is for the volume pad and see if there has been any spikes or something like Nancy, could you share the screen?

Like if you look at this, see, this is a terrible, this is a terrible volume pattern. I mean, there was, there was a sudden spike, right? And if you go to the next slide, I'd say these are some examples of your, you [00:25:00] know, really bad volume patterns and next slide piece. So if you have patterns like this, your repetition is going to be like that. It's going to be, you know, on the bad or the low level, and it's not going to inbox. It's not going to, no matter what high volume you send, it might probably get half of them might get logged and the rest half might go to span and you go to next slide piece. Yeah. Similarly, see the spikes. Aren't really, you know, ISPs don't take like bad patterns and volume spikes kindly your reputation is going to going to get affected.

So that would be something, you know, that would be the, one of the foremost things I'd look at. Absolutely. Right. [00:25:41] Surya: Absolutely. Okay. And besides the volume spikes just coming back to the volume. Why volume spikes are looked at suspiciously by spam filters. It's just because of the pure definition of spam, right? Definition of spam is unwanted emails sent in bulk. So spam filters back in those days would [00:26:00] be looking out for anybody, sending emails in bulk to users who are not subscribed. So I'll give you a classic example, which I tend to give very frequently. So if you have a neighbor who has a car and he wants to sell the car, right? And he sees you have a vacant parking spot, he sends you an email one on one saying, Hey, I want to sell my car. So he has identified an opportunity. He has not sent an email to the entire neighborhood. He has seen that you have a vacant parking spot. That's why he wants to sell it to you. So it makes sense over there.

It's not spam, but if he sent it to the entire neighborhood, it would be deemed as spam. So what Theia said is very important. Don't have a tremendous spike in volumes. I generally recommend marketers more to go beyond 1.5 X of what they have sent in the last 30 days. If you haven't sent anything in the last 30 days, you might have to start from scratch. Okay. And then comes the data pace part. [00:27:00] Whenever your deliverability goes down, whenever your reputation goes down, always look at what changed. Right. Did I target an inactive segment? Did I target users who have the potential to report to spam? And my spam rates high, I will go to Google postmaster and check my spam rates. That's very important. As I mentioned, one of the key factors, deliverability depends on is what kind of segmentation you are incorporating in your email strategy. So for example, if you are a job industry, right job industry, I wouldn't recommend send to the past clickers and openers to get the maximum engagement.

You would be sending campaigns to the recent subscribers because they have. Maximum affinity. They have just signed up and they're looking for a job. Maybe people who signed up 30 days ago, they have already got a job. But when it comes to the fashion industry, you can pretty much target your openers and clickers because over there, the, the tendency to purchase again, the tendency to repeatedly transact is very high.

Okay. [00:28:00] So database is very, very important. Always monitor. What kind of change took place in terms of database targeting then comes content. Did you change the content? Did you introduce any blacklisted URL URLs? Did you use bit links? Bit links are a strict no, I'm rerating again. Yeah. Did you use map tags, right. And then monitor the info reputation is your IP big shared by other brands? And those other brands are spaming. That's why the IP reputation went now. And that's why your domain reputation took ahead. Right? Did you as Ashi already. Is there any DNS failure from the client set because DNS failures cause change in authentication protocols, suddenly you will be not signing DKM on your send domain.

And as a result, boom, all the emails start lining in spa. Postmaster data is very important. Google postmaster gives you delivery errors. Delivery errors also offered when you are sending [00:29:00] emails at a rate at which the recipient can't accept. So typically Yahoo spam filters through an error when you open more than 20 SMTP connections with the recipient, MX of Yahoo. So all these factors need to be looked at very closely. So yeah, [00:29:18] Sadikshya: I was gonna say like Surya said, you know, your segmentation is really important and the one other thing that you need to look at when you look into segmentation as. If you have, like how many, if there has been any unusual activity in the signups, like if you are not a confirmed optin or double optin, you know, customers, then there might be some bot signing in, into your you know, into your brand or into your database.

So in that, so I just look into your welcome pattern in that, in that sense. Welcome, welcome campaigns are really, really important, not just to onboard the customers, but they have, it has some other additional benefits as well. For example, if you have a welcome campaign in place, you are [00:30:00] from the get go. You are basically suppressing all the invalid users, you ID fine and expressing the users, right? That's so that says you won't be having any invalid users in your database. If you have, or you won't have it in, you know, a large quantity. If you have a welcome campaign going on at the same time, you can also monitor the patents of your welcome campaign to see if there there has been need.

Suspicious signup activity on the website. Like if you are running some program, which is for which the expected goal is to get better signups, you know, get better, you know, get more signups then. Yeah. Obviously you would expect a increase in your welcome volume, but on a normal day, it's it? It is, it is expected to follow a regular pattern, right? So that's, as if there has been a irregular pattern, that's also something you should, you know, flag as suspicious and identify it. That can also be, it could also be leading, you know, bots or stand traps into your system. And that could also be something [00:31:00] which can cause your issue. And at the same time, at the same time, the other thing ISP look into is history. The performance history example, if you are a, if you have a sketchy history going back, you know, several months, then even a very small problem can get you into, you know, Into a drop in reputation and you can start going to spam, but then if you have a good history going back a long time, so you have a ear worth or two years worth or worth of a good sending history, you know, good engagement, bad and stable reputation, everything.

If you have been doing everything right for the past two years, then maybe one small mistake and small error that might get excused at this on. Or when you, or even if you have an issue, when you create a remediation request, you can tell them, Hey, this was a mistake. This is what happened. So could you please help us remediate? Right. So those requests will get accepted. I would say, never, [00:32:00] never, ever lie in a support ticket, right? If you are raising a support ticket to Microsoft, and if you tell them, if you lie to them, I mean, they have your history better than you. There's no point lying to them, right? Just so that's its history is really important, but if you have a good history, then. Small, some small errors can get excused, but if you have a sketchy history, then that you are going to get flagged in very minor errors also.

[00:32:26] Surya: Right? Absolutely. I agree with you and with Microsoft on the side note, even when you are writing the truth, sometimes they don't uh, whitelist your, so forget the forget. Like if you, if you lie, you are doomed and Sadia has absolutely bang on when you have a sketchy history. One bad day can drown you and coming back to welcome CDs, right? What Sadia was saying, it's the most organic way of filtering out bounces. Don't go in for data cleansing tools, don't buy data cleansing tools.

You're wasting your money. Just introduce a welcome [00:33:00] series and everything will be said. And in the us people use welcome series, not only to build brand affinity, not only to figure out bounces, but also they act as a preference center. Welcome C. Sets the tone of email marketing for any user. So you ask the user, how many times the user wants to receive the email? What kind of content, if you're on newsletter website, what ki kind of content the user is interested in out of all the topics that you have and at mail model, you can pretty much insert this within a widget through a drag and drop editor within the email itself. And you can collect user preferences in a very sophisticated fashion

[00:33:39] Sadikshya: adding to that right. And one another thing that is, if you have subscribed to any inbox, placement are measuring tools like EDS or a written path, right? So you, you should also monitor your inbox placement, right? Sometimes the changes in open, you know, open metrics or others might not be that visible, but if you see inbox placement getting, [00:34:00] you know, getting dropped, or if you see it's been. Slightly, but slowly, if it's dropping in a few days, then that might be something you should consider and you try to fix before it hits your reputation. So for the gradual decreases and all, you should also monitor your inbox placement. So marketers over here you, can you name a few such tools which help us track in and boxing? Sure. It's E data source and return path. Those two are on that though. Having said that inbox place in data is very essentially well monitoring, but you shouldn't take inbox, placement data as a, you know, it's not, it's not the absolute truth because it's basically giving you results based on certain number of family users or feed users.

So you should also data. Yeah. So it's extra data based on a small size of sample users. So when you're looking at inbox, ment data also look into how big your sample sizes. And also pair it with your performance metrics that you see, right? So then one, [00:35:00] you'll get a better idea of how you are performing and at the, and at the same time, postmaster data is, is just the opposite of this one because postmaster your Gmail, postmaster data is basically exact what Gmail is telling you, right? It's not postmaster data is basically how Gmail sees you performing. It's not like they have, you know, from the get, go about your score and based on your performance, they're lower or, you know, this one or their increase or lower your score. It's basically they give you a score based on how you perform. So it's an indication of your, like, your reputation is basically indication of how well you are performed at its reputation. Data is really important,

[00:35:44] Surya: right? So inboxing data is more like going to a doctor and postmaster data is the taking at the standard. yeah, the scan results recommended by the doctor is how I would put it.

[00:35:57] Sadikshya: So I guess if we have [00:36:00] to sum it up, then I would say check for your authentication results. Check. If there has been any change in the, in your setup. Like, you know, if you have certainly moved from shared setup to a dedicated setup or vice versa and see if there has been any change in IPS or basically all the setup and authentications look for your volume spikes. Look for your unusual target segments. Look, look for unusual signups. Look for the reputation data and your inbox, placement data, check your performance history, you know, going back 15 days and weekly performance and monthly performance history going back six or so months. Check your inbox, placement data, check your postmaster data and check your SGML content. See if you have made some changes in TML content, which was something unusual that you have been doing, you haven't been. And that says that might have triggered a span filter. Right. So just these are the things we should look for. Right? Absolutely. [00:36:56] Mansi: Thanks for summing it up for now. [00:37:00] And yeah. So let's just take up this question that has come up from Navajo.

It's about open rates. Can you, any of you take it up? [00:37:06] Surya: So firstly, you know, as mentioned by other steps that you should follow, right? Look at your postmaster, look at your postmaster and see if there has been a different reputation, analyze the root cause by the different reputation. Have you stopped signing authentication? Have you changed the content? Did your promo emails affect your transactional emails because you send the emails to some inactive database, right? So reputation is. And if your reputation hasn't changed, check the tab, placement check. If your emails test all the emails, the transactional emails, all the use cases. And also the promo emails see where the oil campaigns were landing and where the current campaigns are landing. So actually, if you want to add onto that,

[00:37:57] Sadikshya: Yeah, so I would say first of first things, [00:38:00] first, I would say put your transactional emails in a separate domain in IP and put separate out your promotional emails, everything non transactional, in a separate set of all together. Like your transactional emails are, they're very important to maintain your, maintain a trust with your customer, right? Between the customer brand. Like if I purchase something I'm going to want to want to see the receipt, like my payment has been successful. The purchase history is if it's been shipped or.

You know, at the same time, if you are banking vertical, obviously I'm expecting a statement. So not sending those very essential emails that can, you know, that can lead to some suspicion, you know, from the customer and on like, it could lead to some trust issues. That's it affecting your like put your transaction email completely into a separate separate domain and you know, make sure it's properly set up and you are doing it right. And don't pair your don't promote your promotional [00:39:00] content in your transaction emails, right? That just, that's basically some sort of cheat or hack. That's like, that's not what I wanna see if I want a payment receipt, I'm not going on to see product recommendations. In my payment receipt, you can send a separate email.

Like, you know, you could have a flow campaign, which is basically a post purchase, right? A post purchase flow campaign where after X number of days or X number of hour X, number of days after my purchase, you send me. Yeah, similar recommendations. You know, like people who bought this product also bought this. See if you're interested, something that you could have an entire post flow post purchase flow can be, but I'm definitely not going to want to see your recommendations on my receipt. So first thing is, are separate out your transactional and non transactional emails completely at such your non transactional emails don't impact your transactional ones.

And, and again, like SU said, if you have, if you have started to see a decline or if you see any issues, the first [00:40:00] thing is to debug it and find out where the issues came from. And at one point did it started, if you see a slow decline, then you have to figure out at one point, at what point did they start a decline, right? If they started to gradually decline over the past six months, and you went from 10 to 9.8 to 9.5 to 9.2 and like, and so on. So that is also a decline might not be visible in the first first look. Right. But if you, if you look into your weekly and monthly data, then you'll be able to see the trends.

That's why I said history is very important. Check your engagement history and see what's going. And then pad it with your reputation data, see if on those days how the reputation trended and how your inbox placement trended, and just figure out what you did wrong and, and then, you know, fix it from there. Right. And I would say one basic thing that works for everybody is if you start to see decline [00:41:00] then for the next next seven days, just into your, you know, most active users and lower your sense frequency. Absolutely. Absolutely. So so that you, you know, like this is a very real world scenario. What you're talking about ideal world scenario.

[00:41:14] Surya: I always assume the promotional emails are sent to. Good databases in that case, if, if your database is really clean, because I've seen brands, enterprise brands like Micah and , you know, like they send emails they use the same domain to send transactional emails and same domain to send promotional emails over there. Why it is not a bad practice is because they maintain a solid email marketing strategy. They never, they never like, you know, send only to inactive users. There are two sides to the go ahead. So the other side is you can piggyback on your transactional email, open rates and sustain the reputation of your domain in, but that is only if only if you're targeting the best users in the mailing list. If you're not confident about your [00:42:00] database, which is 99.9, 9% of the cases do as what SEIA just mentioned, separate out, the two pipes, transactional will be one pipe and promotional will be another.

[00:42:12] Sadikshya: I would say there's a flip side to it, right? When you pair your good customers with your transactional setup, what do you do about your less active or inactive users? They're not going to sustain a, a reputation entirely by themselves. So that's just, there's also a flip side to it. I would say that's. Yeah. So I would say for the long run, I mean, for the short, short term, it obviously works. Like if you suddenly send some promotional campaigns through your transactional center, it's obviously going to get good placement and it's obviously going to get better opens, but that's just short term, right? That's I mean, if you send me a transaction promotional email from my transactional, from your transactional setup, I might get it in the primary folder, but that's not where I want it. That's not where I have been seeing it for the past, whatever, so [00:43:00] months. So I might appreciate, I might not appreciate it. I might unsubscribe where to report to this spam or move it back to promotion, which is. Which has, you know, even larger impact in your tap placement. So at such, I would say, you know, following best practice is always the key. Like as long as you follow best practice, you can really sustain it. That's sustainable hacks are never sustainable. Right? Absolutely, absolutely agree. Yeah. Uh, That moving on

[00:43:26] Mansi: man, I think uh, we, we can take up this follow up question from number, if you think uh, we can answer this one here. So he is asking what are the flows, if you started to see it dip in the open rates of the email flows.

[00:43:40] Sadikshya: Okay. I, I would say, I would say, look into your sending frequency. Like you try, try moving about your after, like in your flow campaigns, try moving about the deals. Like, if you're sending up your email, like three hours after certain event, try moving the hours and see how it performs and put a frequency [00:44:00] gap to it.

[00:44:01] Surya: Right? Another very key aspect is monitor. If you have introduced any other nudge. For example, I was assisting a brand. They were sending abandoned card emails within 30 minutes, and they suddenly went live with WhatsApp. And within 15 minutes of abandoning the card, they started sending WhatsApp modules. So all the conversions shifted to WhatsApp. So that is very important. And secondly, also, again, monitor your domain reputation. See if you have had any tip over there, you two, maybe you are sending promotional campaigns from the same domain. So monitor that.

[00:44:38] Mansi: Cool. I think we can take up the other questions that have come up later. Once let's let's finish up our uh, set of questions that we've had for you. Let's come to our last point, which is where we can talk about the evolution of spam . [00:44:54] Surya: Absolutely. So like I I'll tell you guys a very interesting story. So in the [00:45:00] early two thousands when spam originated spam filters was simply rule based. That's mainly because spam originated from a few set of servers and the anti-spam guys knew about those servers. So spam would be any email sent from those particular ranges of IPS. So they would simply make a rule in the spam filter that if the IP belongs to this range, block the email, and then anti-spam versus spam has always been a cat and mouse game. Right? So spam started thinking what if I start relaying my email through some remote server? So if someone from Nigeria was spamming, they would route the email through the. To your recipient S and the emails would be accepted because the spam filter read the center IP as IPS belonging to UK, right? So that's when anti span guys sat down together and they decided let's make up protocol called the center policy framework. That's how you have SPF today. It basically authenticates the center server from [00:46:00] which the email is being sent from then to now it, it, it then transitioned to Basian filter algorithms. And now we have aim algorithms, which, which is, you know, like beautiful, the page, email works. It's it's next level of AI and ML, right? They, they read thousands, billions of emails per day, and they're able to decide if the user is supposed to receive it in inbox plus spa. Okay. And how I would say even other ISPs spam filters. Changed over the years like outlook Yahoo outlook is one of the toughest pan filters to inbox. Sometimes it throws a lot of false negatives, but, you know, following the best practices is always the key. For example, some spam filters like outlook and iCloud require you to sign a dedicated PDF on your IP address. That is also very each, each and every spam filter has their own [00:47:00] works. Ion.

[00:47:03] Sadikshya: Yeah. I'd say spam filters are, you know, they have been evolving very, very quickly, like in the beginning, like couple of years back when we got the postmaster, there was like a very new thing. I think it was 20, 50, 20 16. So we suddenly had the reputation data. Right. So you could, in a, in the matter of two days, you could tell where your domain is trending, how, where it's going. Right. And then a couple of months back. I mean, Yahoo was, I mean, personally Yahoo is somewhere, but it was easy to spam. Like my Yahoo mailbox used to be flooded with all sorts of spams. So I'd actually moved out of using my Yahoo mailbox and I'd only open it once every, every few weeks or every few months to just clean it up. Right. A few months back, I opened my Yahoo mailbox and lo andhold. They were like, there was just Uber receipts. That was, that was a very, very pleasant surprise. So that, that tells you how they're evolving, right? How [00:48:00] ISP filters are working on changing their algorithms, how they're working actively to prevent spam. So that's suspicion of so ANSYS, we should. So at ANSYS, we should also, as a marketers, we should also evolve our strategy. We should also change and see optimize and evolve the same way. And one, some few things you need to monitor is, again, like I said, history is important.

Like yeah, Hotmail, Microsoft is one of the toughest inbox. If you have irregular pattern. Not just put in, put your emails in span, they'd actually outright and block you. I mean, I have a few customers who send who basically segment based on their reasons, right? So, which is a good practice. They're actually segmenting and sending very reselect content, but they get blocked very frequently because Microsoft sees very at pattern because based on each reason, but in those cases, because they're actually sending Liz legitimate emails, when you raise a support request, they get remediate. But at the same time, if you have that pattern for no good reason, then that's not going to, [00:49:00] you know, that's not going to fix, get fixed anyway, at, at Gmail, you can, you can literally see, I I've, at many cases, I've literally seen a customer accident. He blasted a high volume. So the inbox placement started off with a hundred percent.

Like it had been in, you know, in whatever last six, six or so months. And then as the volume kept on going, like, you know, based on the throtle as it kept. Then it's slowly, DIPP from hundred to seven, 80 from 80 to 70 to 60 to 30%. You can literally, you know, I wouldn't say live, but you can at real time, see how you inbox.

This is dropping at female. And same with Yahoo, Yahoo, Yahoo has recently changed, just filter and make made it much stricter. And though I'd have to say, it's, it's very commendable that when they have tightened up their security, the actual actual customers, you know, the actual CRM centers aren't really affected with that. So I think which is great. And [00:50:00] Yahoo has a slightly lower acceptance rate compared to Gmail. So that's as the frequency, you know, the limits at which you are sending at Gmail will not be accepted at Yahoo at the same time, the number of IPS we have used to send Gmail volume, that those number of IPS might not suffice when you are targeting the same amount of, you know, the higher amount of Yahoo. So, those are also some factors you really keep in mind.

[00:50:28] Surya: In fact, if you're monitoring your SMPT servers, please ensure you're not opening more than 20 connections, SMTP connections at the same time with Yahoo, otherwise it throws an error. So that's a takeaway. I think, man the audience is getting impatient. We have a couple of questions over here.

[00:50:44] Mansi: Let's take one from the name first. Yeah.

[00:50:48] Surya: Yeah. you go first.

[00:50:51] Sadikshya: Oh, okay. If your welcome series is getting 20% bounce rate, that that means there's something. I'm absolutely wrong with how people are signing up, [00:51:00] like on your welcome pay, like on your signup pace, you have to have some requirements, even if you don't have confirmed opt in, like our double opt in, you should at least have capture in place. Right? 20% bounce rate is like really, really high. Well, you know, higher bounce rate in welcome series is expected. At least see the first campaign it's expected, but 20% is very high. You should add. You should either have confirmed optin in place or you should at the very least have capture in place. Right?

[00:51:28] Surya: Capture is a capture is a very good idea. And also you can have a checkbox asking the users to check if they want marketing emails, because at times people given wrong email addresses typos in the fear of being bombarded by marketing emails. So that is also one aspect to look at.

[00:51:47] Sadikshya: And that's also a GDPR requirement. If you are based in Europe. [00:51:53] Surya: Yep.

[00:51:53] Mansi: Let's take up the next question from Monick advice for companies [00:52:00] expanding to engine market. What's an ideal open in C you can actually check the benchmark report. I, well, I think Nancy has posted it somewhere. We just check it on the chats. Yes. So we have a benchmark report which basically for each reason and each vertical. And for how, for the level of personalization you have used, there's a benchmark report by mortgage. You can check it out on the chat, [00:52:27] Surya: I think. Yeah, I think [00:52:29] Mansi: S there do you wanna add here or anything? Should we go to the next question?

[00:52:35] Surya: No, absolutely. I, I would also benchmark report. But one thing to note is, as you have rightly asked in India, the benchmarks are different industry wise. The benchmarks are different. It depends on which industry you are, from what data set you are targeting. And obviously, you know what kind of infrastructure you are maintaining, but when it comes to benchmarks, do refer to the benchmark report by more engaged.

[00:52:58] Mansi: Yes. I've put [00:53:00] the benchmark report in the chat and I'll put again once we chat. Great. So I think we are almost at the end of this chat here. Let's see if anyone has any more questions, please feel free. We have one more minute. We can take up one or two questions max, but yeah, please feel free. Or I think we can just wait for. Surya and Sadikshya. Do you have anything like a parting note or anything to add your for the audience? [00:53:30] Surya: Parting note would be, there is no one size fits all kind of solution. Always treat each and every customer as a separate detective case that you're solving as if you are in a clientele and you are a cop, always keep that in mind. So that's your parting notes from you? I

[00:53:50] Sadikshya: absolutely agree with you. In fact, the reason I, I know I like my job so much is because every case, every customer, every client, every issue they're so different. I mean, [00:54:00] you might have repetition issue for two customers at the same particular, but their underlying problem and how you solve them can be entirely different. Right? So it's, it's quite exciting. You're always learning new things and everything can affect it. Though, I would say one thing that works for everybody is follow the best practices hacks, and they work for a short time. If you have issues, okay. Change the IP, right. That might work for a day or two, or they might even work for a week or so. But what about after that as a brand loyalty is important billing, you know, billing trust with your end users is important. So I would say always follow the best practices, listen to your users' needs. Look at the Excel for spas

and sorry. And yeah, Excel for spas and fixes are for market years. Always remember that and the other thing is with all the new laws, you know, every country is strengthening their privacy policies and span filter is getting all [00:55:00] stricter. They're not, they're not, I mean, obviously they pose a issue if then you have to change things best than that, but they're not. Affecting you as a marketer who you really wanna target where your conversion comes from is the interested users. The privacy policy is just ensuring that you actually target to just them and not bombard the people who don't want losing this. Right? So at this, you should look at them as, oh, they're basically making you follow the best guidelines. So I'd say just, you know, follow the best practices. Absolutely good.

[00:55:34] Mansi: Let's let's wrap up here and it was great talking to both of you. We had a lot of insights to take away. We also share the recording with everyone and so that, you know, they can use that and learn and keep optimizing their delivery deliverability through this video. Thank you guys. Thank you for joining us today. You soon, next time, byebye. Yeah, it was great talking. Bye. I.[00:56:00]